Creating maps in R

Toolkit

library(tidyverse) # all of our usual tools
library(ggmap)     # for spatial viz.
library(viridis)   # for color scales

Oklahoma earthquakes

• The number of earthquakes in Oklahoma has increased in the past decade

• We will examine data obtained from the U.S. Geological Survey (USGS) on earthquakes in Oklahoma since 2000

ok_quakes <- readr::read_csv("OKquakes.csv")
str(ok_quakes)

## Classes 'tbl_df', 'tbl' and 'data.frame':    5204 obs. of  9 variables:
##  $ time     : POSIXct, format: "2015-12-31 20:31:14" "2015-12-31 11:35:26" ...
##  $ latitude : num  35.8 36.6 35.7 36.8 35.7 ...
##  $ longitude: num  -97.6 -98.8 -97.4 -97.8 -97.4 ...
##  $ depth    : num  6.48 6.29 5.97 5.41 6.18 ...
##  $ mag      : num  2.6 3.2 2.5 3 2.8 2.5 3 3.2 3.1 3.3 ...
##  $ place    : chr  "16km NNW of Edmond, Oklahoma" "24km SSW of Alva, Oklahoma" "6km ENE of Edmond, Oklahoma" "4km NW of Medford, Oklahoma" ...
##  $ year     : int  2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 ...
##  $ month    : int  12 12 12 12 12 12 12 12 12 12 ...
##  $ type     : chr  "very minor" "minor" "very minor" "minor" ...

Displaying spatial data with ggplot2

ggplot(data = filter(ok_quakes, year == 2015), aes(x = longitude, y = latitude)) +
  geom_point()

The ggmap package

• Context is important, we need more than just the points!

• ggmap allows us to grab this context from google or other sources

• By default, get_map, downloads a terrain image from Google maps

# Download the requested map
ok_map <- get_map(location = 'oklahoma', zoom = 7)

## Map from URL : http://maps.googleapis.com/maps/api/staticmap?center=oklahoma&zoom=7&size=640x640&scale=2&maptype=terrain&language=en-EN&sensor=false

## Information from URL : http://maps.googleapis.com/maps/api/geocode/json?address=oklahoma&sensor=false

# Display the map
ggmap(ok_map)

1. Look at the help file for get_map and read about the location and zoom arguments. Do the following:

• Use location = 'lawrence university' to get a map of the surrounding area. Try difference values for zoom.
• Choose a city and use Google maps to find the longitude and latitude of a landmark. Specify the coordinates in the location argument and produce a map of the surrounding area. Again, try different values for zoom.

There are other options aside from a terrain map.

2. Look at the help file for the get_map function and read about the maptype and source arguments. Specify different maptypes and sources. How do things change?

Now that we have our map of Oklahoma, we can add points to it:

ggmap(ok_map) +
  geom_point(data = ok_quakes, aes(x = longitude, y = latitude, color = year))

A better idea might be to facet the plot by year. In order to to this we need to add the base_layer arugment to the ggmap function:

ggmap(ok_map, base_layer = ggplot(ok_quakes, aes(x = longitude, y = latitude))) +
  geom_point(size = .8) +
  facet_wrap(~year, ncol = 5)

3. Add color to the facetted map based on the type of earthquake. Be sure to use an appropriate color palette. Do any patterns emerge?

4. The file CAquakes.csv contains data on earthquakes in California from January 1, 2010 through the end of 2015. Use ggmap to plot the earthquakes that occurred in California during this time, similar to what we did for Oklahoma above. Do any patterns emerge?

Heatmaps

ok2015 <- filter(ok_quakes, year == 2015)

We can create heatmaps using stat_density2d in place of our geom_point layer:

ggmap(ok_map, base_layer = ggplot(ok2015, aes(x = longitude, y = latitude))) +
  stat_density2d(aes(fill = ..level.., alpha = ..level..), size = 0.01, 
    bins = 16, geom = "polygon") +
  scale_fill_viridis() +
  scale_alpha(guide = FALSE)

5. Create a heatmap of California earthquakes in 2015.